Preclinical Evaluation of Zn(II) Self‐Assemblies with Selective Cytotoxic Activity Against Cancer Cells In Vitro and In Ovo

Abstract Dipodal pyridylthiazole amine ligands L1 and L2 both form different metallo‐supramolecular self‐assemblies with Zn2+ and Cu2+ and these are shown to be toxic and selective towards cancer cell lines in vitro. Furthermore, potency and selectivity are highly dependent upon the metal ions, ligand system and bound anion, with significant changes in chemosensitivity and selectivity dependent upon which species are employed. Importantly, significant anti‐tumor activity was observed in ovo at doses that are non‐toxic.


A. Chemistry Studies General Information
Chemicals were purchased and used without further purification. 1H and 13 C NMR spectra were recorded on a 400MHz Bruker Avance DP X400.Mass spectra were obtained on an Agilent 6210 TOF MS for the organic species with the metal complexes run on a Bruker MicroQTOF LC.
CAUTION: perchlorate salts are potentially explosive and should be treated with due care.Those complexes described below which were isolated as perchlorates were only prepared in small amounts (5 -10 mg) and we had no problems with them.
Ligand Synthesis. 1   To a 100 mL RBF charged with putrescine (0.35 g, 3.9 mmol) and MeCN (25 mL) was added a solution of benzoyl isothiocyanate (1.3 mL, 1.59 g, 9.8 mmol) in MeCN (25 mL) slowly over 30 min with constant stirring.The reaction mixture was then stirred at RT for a further 12 h during this time a heavy white precipitate was formed.The resulting mixture was the added dropwise, whilst stirring, to deionised water (100 mL) to yield a cream precipitate which was then isolated via vacuum filtration.

Synthesis of L 2
To a 250 mL RBF was charged with ethylenediamine (1 mL, 0.9 g, 15 mmol) and MeCN (50 mL) was added a solution of benzoyl isothiocyanate (5 mL, 6.1 g, 37 mmol) in MeCN (50 mL) slowly over 30 min with constant stirring.The reaction mixture was then stirred at RT for a further 12 h during this time a heavy white precipitate was formed.The resulting mixture was the added dropwise whilst stirring to deionised water (300 mL) to yield a white/yellow precipitate which was then isolated via vacuum filtration.The solid was then suspended in MeOH (50 mL) in a 100 mL conical flask and sonicated (5 min).The white/yellow suspension was then once again filtered under vacuum to yield a fine white solid which was washed with further portions of MeOH (3 × 10 mL) to give the dibenzoylated dithiourea as a fine white powder (3.5 g, 60%). 1   In a 100 mL RBF was combined the benzoylated dithiourea (1 g, 2.6 mmol), deionised water (30 mL) and a magnetic stir bar.A solution of NaOH (0.63 g, 15.8 mmol) in H2O (15 mL) was then added to the reaction mixture whilst stirring at 60 °C.After 12 h the resulting colourless solution was then cooled to RT during which time a fine white precipitate was formed.The precipitate was then collected via vacuum filtration and then washed with deionised water (3 × 5 mL) giving the thiourea as a fine white powder (0.43 g, 93%).The 1 H NMR gives four broad signals in the aromatic region and two broad signals at ~ 3.5 ppm.It is suspected that intra-molecular hydrogen bonding is inducing broad peaks in the 1 H NMR. However, both the 13 C and ESI-MS are exactly as expected. 13C NMR [400 MHz, (CD3)2SO]: δC = 183.9(C=S), 43.7 (CH2).ESI-MS m/z 179 (M + H + ).HR ESI-MS found 179.0415 C4H11N4S2 requires 179.042 (error 2.61 ppm).Synthesis of L 2 .A 25 mL RBF was charged with dithiourea derivative (228 mg, 1.3 mmol), EtOH (10 mL) and magnetic stir bar and the mixture was then heated to 60 °C and to this added α-bromoacetyl pyridine (540 mg, 2.7 mmol) in a solution in EtOH (2 mL) solution whilst stirring.Heating was continued for a further 8 h during which time a heavy yellow precipitate was formed.This was then isolated via vacuum filtration and the filtrand was then washed with portions of EtOH (3 × 2 mL) to give the protonated product.The free-base ligand was then isolated by suspending in concentrated ammonia (15 mL) for 24 h.The resulting colourless suspension was then sonicated (5 min) and was filtered under vacco and the filtrand washed with portions of deionised water (5 × 2 mL) and EtOH (2 × 2 mL) giving the product as a fine off-white powder (402 mg 81%). 1 H NMR (400 MHz, (CD3)2SO) δ (ppm) 8.55 (d, J = 4, 2H, H-6 py), 7. overlapping,4H,7.79 (dt,J = 7.6,1.6,2H,7.31 (s,2H,tz),7.26 (ddd,J = 7.4,2.6,1.2,2H,3.60 (d,J = 2.5 Hz, 4H, -CH2-).13 C NMR (100 MHz, (CD3)2SO) δ 169.0, 152.9, 150.7, 149.7, 137.4, 122.8, 120.8, 105.4   Synthesis of [(L 1 )2Zn2(H2PO4)](ClO4)3.To a solution of Zn(ClO4)2⸱6H2O (9.1 mg, 0.025 mmol) in MeCN (1 ml) was added a suspension of ligand L 1 (10 mg, 0.025 mmol) in MeCN and the reaction gently warmed and sonicated until a clear solution had formed.To this was added a solution of Bu4NH2PO4 (3.9 mg, 0.012 mmol) in MeCN (0.5 ml) and dichloromethane was slowly allowed to diffuse into the solution resulting in the formation of colourless rod-like crystals after several days.

Synthesis of [(L 1 )4Zn5(NPP)5](ClO4)3.
To a solution of Zn(ClO4)2⸱6H2O (9.1 mg, 0.025 mmol) in MeCN (1 ml) was added a suspension of ligand L 1 (10 mg, 0.025 mmol) in MeCN and the reaction gently warmed and sonicated until a clear solution had formed.To this was added a suspension of disodium 4-nitrophenylphosphate (4.9 mg, 0.012 mmol) in MeCN (0.5 ml) and the rection warmed and sonicated until all the anion dissolved.Ethyl acetate was slowly allowed to diffuse into the solution resulting in the formation of pale-yellow faceted prismatic crystals after several days.Filtration and washing diethyl ether (1 ml) gave pale-yellow crystals which lost solvent rapidly (yield = 42%).

Synthesis of [(L 2 )2Zn2(H2PO4)](ClO4)3.
To a solution of Zn(ClO4)2⸱6H2O (9.8 mg, 0.026 mmol) in MeCN (1 ml) was added a suspension of ligand L 2 (10 mg, 0.026 mmol) in MeCN and the reaction gently warmed and sonicated until a clear solution had formed.To this was added a solution of Bu4NH2PO4 (4.2 mg, 0.013 mmol) in MeCN (0.5 ml) and dichloromethane was slowly allowed to diffuse into the solution resulting in the formation of colourless plate-like crystals after several days.

Synthesis of [(L 2 )4Zn5(NPP)5](ClO4)3.
To a solution of Zn(ClO4)2⸱6H2O (9.8 mg, 0.026 mmol) in MeCN (1 ml) was added a suspension of ligand L 1 (10 mg, 0.026 mmol) in MeCN and the reaction gently warmed and sonicated until a clear solution had formed.To this was added a solution of disodium 4-nitrophenylphosphate (4.9 mg, 0.013 mmol) in H2O (1 ml) and the rection warmed and sonicated until all the anion dissolved.The solvent was then allowed to evaporate which deposited colourless plate-like crystals after several days.Filtration and washing diethyl ether (1 ml) gave colourless crystals which lost solvent rapidly (yield = 35%).

Crystallography
Single crystal X-ray diffraction data was collected at 150(2) K (100 K for [(L 1 )4Zn5(O2NC6H4PO4)4] 2+ ) on a Bruker D8 Venture diffractometer equipped with a graphite monochromated Mo(K) radiation source and a cold stream of N2 gas.Solutions were generated by conventional heavy atom Patterson or direct methods and refined by full-matrix least squares on all F 2 data, using SHELXS-97 and SHELXL software respectively. 2Absorption corrections were applied based on multiple and symmetry-equivalent measurements using SADABS. 3  [(L 1 )2Zn2(ClO4)2](ClO4)2.The structure contained positionally and rotationally disordered dichloromethane solvent molecules which were modelled over two positions using the PART instruction and one of the solvent carbon atoms was restrained using ISOR.
Two of the solvent molecules had their carbon chlorine bond lengths restrained using DFIX and some of the solvent molecules were restrained with DELU, SIMU and ISOR.
[(L 1 )4Zn5(O2NC6H4PO4)4] 2+ .Only a quarter of the molecule (comprising one ligand, one NPP dianion and one and one quarter Zn 2+ atoms) was contained in the asymmetric cell and the remaining 3/4 of the assembly formed by symmetry operations.The structure contained whole molecule disorder with the ligand and NPP dianion modelled over two positions using the PART instruction.Due to the extensive disorder some bonds in the disordered ligand had to be constrained with DFIX instructions as well as AFIX 66 for the 6-membered rings.Further to this ISOR was used globally for the carbon, nitrogen and oxygen atoms.The structure contained diffuse electron density, which despite attempts, could not be modelled and was removed using the solvent mask facility in Olex2. 4The solvent mask removed a total of 284 electrons per pentanuclear assembly which corresponds to 2 perchlorate anions 2 molecules of acetonitrile and 3 molecules of ethyl acetate.Despite removal of extensive number of solvent molecules with the use of the solvent mask function satisfactory data was obtained.
[(L 2 )2Zn2(H2PO4)2](ClO4)3.Only a half of the helicate molecule was present in the asymmetric unit and the remaining half was formed by symmetry operations.The structure contained a rotationally disorder perchlorate molecule and the oxygen atoms were modelled in two positions using the PART instructions and constrained using DELU, SIMU and ISOR.These three restraints were also used on one of the dichloromethane solvent molecules.
[(L 2 )4Zn5(O2NC6H4PO4)4](ClO4)2.Unlike the L 1 derivative the whole molecule was present in the asymmetric unit cell.However, two of the ligands had disorder associated with the ethyl spacer unit and one of the thiazole rings and these were modelled in two positions using the PART instructions and constrained using DELU, SIMU and ISOR.Also present was solvent disorder that despite attempts could not be successfully modelled and as a result the diffuse electron density was removed using the solvent mask facility in Olex2, resulting in voids in the crystal structure. 4The solvent mask removed a total of 293 electrons in the asymmetric unit which corresponds to one perchlorate anion and eleven molecules of acetonitrile.Analysis of the complexes by 1 H NMR was carried out in CD3CN/CD3OD (2:1) as this solvent system dissolved all the complexes although precipitates did form in the Na2NPP system.The            Collection apart from the p53 +/+ and p53 −/− isogenic clones of HCT116 human colorectal adenocarcinoma cells were a kind gift from Professor Bert Vogelstein (Bunz et al., 1998).The cell lines were maintained as monolayers, at low passage and in antibiotic free media.HT29, DLD-1, HCT116 p53 +/+ and HCT116 p53 −/− are all human colorectal adenocarcinoma cell lines; PSN-1, BxPC- In ovo efficacy studies: In ovo experiments using the chick embryo model were performed in accordance with UK legislation and ethical guidelines with all embryos being humanely terminated on day 14 of chick embryonic development (E14) as stipulated by the UK Animals Scientific Procedures Act 1986 (amended 2012).E14 represents two-thirds of the chick embryo gestation period and up to, and including, day E14 the chick embryo is classified as a non-protected, 3Rs compliant model with no animal license or home office approval being required.
Fertilised Shaver Brown hen eggs were purchased from Medeggs Ltd and stored at 14 0 C for up to a week.To commence embryonic development (day E0), eggs were incubated on their side in egg trays at 37.8 º C and 45% humidity in a specialized poultry egg incubator (Brinsea OvaEasy 380) for 3 days (with the upwards side of the egg labelled in pencil to orientate for E3 windowing).Incubator shelves were set to alternately tilt 45º from a horizontal position every 45 minutes.On day 3 (E3), the wide base of each egg was gently pierced and 7 mL of albumen were removed to lower the chorioallantoic membrane (CAM) from the eggshell before windowing.Eggs were windowed by piercing a small hole on the labelled side of the egg, applying a 3 cm piece of invisible tape over the eggshell area to be windowed, and cutting using fine scissors a small three sided window (~2cm by ~1 cm by ~2 cm) in the eggshell (Barnett et al., 2022).Windowed eggs (E3) were further incubated without tilting at 37.8 º C and 45% humidity until day E7.
On day E7, eggs containing viable embryos were implanted with tumour cells.PSN-1 pancreatic cancer cells growing as monolayers and in logarithmic phase growth were harvested from cell culture flasks by standard trypsinization, counted, pelleted at 200 g for 5 minutes and resuspended in a small volume of sterile PBS to generate a concentrated cell slurry.For efficient engraftment of tumour cells, a small area of the highly vascularized chorioallantoic membrane (CAM) was dried with sterile gauze onto which a 4 mm silicone ring was carefully placed.4 x10 6 PSN-1 cells were pipetted inside the ring complexes, the sole exception being against the H460 lung cancer cell line for which the L 1 copper complex showed superior cancer selectivity (Figure S2.2).
For the zinc L 1 and L 2 complexes, the effects of their pre-incubation with H2PO4ˉ, PhOPO3 2 ˉ or O2NC6H4OPO3 2 ˉ anions before cell exposure, on their potency and selectivity towards selective cell lines was evaluated (Figure S2.3, Figure 7 for H2PO4ˉ).Activity of zinc L 2 complex against the ARPE-19 non-cancer cells was modestly increased by all three phosphate anions (decreased IC50 relative to [(L 2 )2Zn2] 4+ alone, Figure S2.3) whereas activity towards the tested cancer cell lines decreased or did not change.This resulted in a small decrease in cancer cell selectivity of the zinc L 2 complex through pre-incubation with phosphate anions.In contrast, pre-incubation of H2PO4ˉ and PhOPO3 2 ˉ with zinc L 1 complex decreased its activity against the ARPE-19 non-cancer cells and resulted in a modest increase in the cancer cell selectivity of [(L 1 )2Zn2] 4+ (Figure S2.3d, Figure 7 inset).Whilst the differential effects of the anions on the zinc L 1 and L 2 complexes requires further investigation, this indicates how potency and cancer cell selectivity which varies depending on choice of both ligand and metal can be further differentially modulated by anion pre-incubation.

Figure S1. 20 .
Figure S1.20.Aromatic regions in the HSQC spectra (CD3OD/CD3CN 1:2) of L 2 plus Zn(OTf)2 and 0.5 equivalents of Bu4NH2PO4 showing a number of 13 C signals indicating the formation of a number of different species.

Figure S1. 21 .
Figure S1.21.Aromatic regions in the HSQC spectra (CD3OD/CD3CN 1:2) of L 1 plus Zn(OTf)2 and 1 equivalent of Bu4NH2PO4 showing the reduced number of 13 C signals indicating the formation of one major species.

3 and
MiaPaCa2 are human pancreatic carcinoma cell lines and A549 and H460 are human nonsmall cell lung carcinoma cell lines.HT29, DLD-1, PSN-1, BxPC-3, A549 and H460 cell lines were cultured in RPMI-1640 growth media (Sigma) containing 2 mM L-glutamine, 1 mM sodium pyruvate and 10% foetal bovine serum (FBS).HCT116 (p53 +/+ and p53 −/− ) and MiaPaCa2 cell lines were cultured in Dulbecco's Modified Eagle's Medium (Sigma), 2 mM L-glutamine and 10% FBS.The ARPE-19 human retinal epithelial non-cancer cell line was cultured in DMEM/F12 media (Gibco), 2 mM L-glutamine, 1 mM sodium pyruvate and 10% FBS.Chemosensitivity studies:The response of cells following a continuous 96 h exposure to test compounds was determined using the MTT assay.[(L 1 )2M2] 4+ and [(L 2 )2M2] 4+ complexes (M = metal ion, Zn 2+ or Cu 2+ ) were freshly formed by adding DMSO to the pre-weighed individual components and then mixing by pipetting.The compounds were then further diluted in cell culture media with the final DMSO concentration that cells were exposed to being 0.1% (vehicle control).All cell lines were seeded into 96 well plates at 2×10 3 cells per well and incubated overnight at 37 °C.Following incubation, media was removed and replaced with fresh media containing the test compounds at a range of concentrations.Cells were incubated with these compounds for 96 h after which the media was removed and replaced with fresh media (200μl/well) containing MTT solution at a final concentration of 0.5 mg/mL and cells were then incubated for a further 4h.Media and MTT were then removed, and formazan crystals that had formed were dissolved in 150 μl of DMSO and well absorbance measured at 540 nm.The concentration of test compound required to reduce cell growth by 50% (IC50) was determined from dose response curves.Potency was recorded as the IC50 ± standard deviation for three independent experiments.The selectivity index (SI) was defined as the ratio of the mean IC50 values for non-cancer to cancer cell lines with values >1 representing selectivity for cancer cells as opposed to non-cancer cells.As the mean IC50 values were used to determine SI, experimental error was not determined.Cell cycle studies:For cell cycle studies, HCT116 and PSN-1 cancer cells and ARPE-19 non-cancer cells were seeded in 25 cm 2 cell culture flasks and 24h post seeding, media was replaced with fresh complete media containing 0.015% DMSO (solvent control), 15 µM [(L 1 )2Zn2] 4+ or 15 µM [(L 2 )2Zn2] 4+ .At 72 h, media containing any non-adhered cells was collected from treated flasks and pooled with adhered cells collected by standard trypsinization.Following centrifugation, cell pellets were washed in phosphate buffered saline, gently permeabilized (solution 10, Chemometec), and incubated with 10µg/ml DAPI for 5 min at 37 º C before analyzing cell DNA content (emission detection at 450 nm for double stranded DNA) by image cytometry (NC3000 image Chemometec).Time lapse microscopy:Real time monitoring and microscopic imaging of cells was performed using Axion BioSystems Lux FL microscopes placed directly inside a cell culture incubator.This enabled the effects of [(L 2 )2Zn2] 4+ on the growth and survival of HCT116 cancer cells and ARPE-19 non-cancer cells to be directly monitored over time within the same magnification field of view.Cells were seeded in 25 cm 2 cell culture flasks and 24h post seeding, fresh complete media containing solvent control or 15µM [(L 2 )2Zn2] 4+ was added.3h following the addition of 15 µM [(L 2 )2Zn2] 4+ or DMSO solvent control, cell impermeable DNA binding dye propidium iodide (PI) was added to the media at a final concentration of 4µg/ml for the positive staining of dead or dying cells.Lux FL microscopes were focused on a particular field of view and programmed to acquire images every 30 minutes for 48 in brightfield (to monitor cell proliferation and changes in cell confluency) and in the red fluorescent channel (to detect PI positive staining and monitor cell death induction).
signals corresponding to at least three species and again exchange between these species was observed.However, addition of a further equivalent of H2PO4ˉ a change in the 1 H NMR is observed and now a single complex is the major species.We attribute this behaviour to the ability of [(L 2 )2Zn2] 4+ to bind two phosphate anions and reaction of [(L 2 )2Zn2] 4+ with two equivalents of further dihydrogen phosphate only resulted in a slight change in chemical shift (attributable to the change in ionic strength and decrease in pH) indicating that the complex only binds one phosphate anion.The 1 H NMR of [(L 2 )2Zn2] 4+ gave signals attributable to two species which were again interconverting.Addition of 1 equivalent of H2PO4ˉ changes the 1 H NMR significantly resulting in a complex spectrum with